NBO



 Dear CCLers,
 Approximately a week ago I send an inquire about NBO (Natural Bond
 Orbital) analysis within Gaussian98 having warning messages.
 WARNING:  1 low occupancy (<1.9990e) core orbital  found on  P 3
            1 low occupancy (<1.9990e) core orbital  found on  C 5
  WARNING:  Population inversion found on atom  P 3
            Population inversion found on atom  O 4
            Population inversion found on atom  C 5
            Population inversion found on atom  C 6
            Population inversion found on atom  C 7
            Population inversion found on atom  C 8
            Population inversion found on atom  C 9
            Population inversion found on atom  C10
            Population inversion found on atom  O11
            Population inversion found on atom  S13
 Thank you all for you help.
 I was asked to summarize this so here all all the answers that I got from
 you.
 &&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
 Hi!
         I cannot say much about the warnings you get from G98 in the
 output.
 But if it's not normal, it might be caused by the bug in NBO link (l607)
 which was confirmed to us by Doug Fox three months ago.  Unusually, the
 bug
 appeared only under Linux.  If your G98 works under Linux, and if you
 purcha-
 sed G98 more than 3 months ago, you should consider this possibility too.
         If you have an access to G98 (or G94) on different platforms and
 if
 you get the same output everywhere, it's probably not related to the bug.
 Otherwise, let me know and I'll send you the copy of my correspondence
 with
 Doug Fox which also includes the fix for the bug.
                 Sincerely,
                    Darko Babic
 &&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
 Those WARNINGs are quiet usual. When NBO program prints Natural Atomic
 Orbitals
 (1s, 2s, 3s etc. ),it does on the basis of the energy order or on the
 basis of occupancy. The
 WARNING message "Population inversion" comes when, the orbital
 occupancies
 and energy ordering don't match or don't coincide. suppose, for example
 when one of 3d orbital lies below in energy, but it's population
 (occupancy) is lower than 4s or any of 4p orbitals, then population
 inversion occurs and should print warning massage.
 For core (NAO treated as unhybridized single center core NBO) orbitals,
 which in formal Lewis sense should be filled by 2 electrons. Sometimes due to
 unphysical mixing of core and valence lone pairs, the core occupancy goes
 below the threshold <1.9990e) and should print worning massage
 However, your NBO analysis should be ok.
 Cheers.
 Jamal
 &&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
 I'm quite new to this area but as i heard, an occupancy
 above 1.99 should refer to a core orbital. Your ones are
 probably  a bit lower than the usual ones (i have no
 experience in it) but i think they might be correct.
      However, if you calculate those occupancies to
 determine which orbitals to use in a multi-reference
 calculation, they might not be reliable: that is what the
 warning is for (in the non-dynamic correlation, those
 determinants/configurations will dominate that have
 orbitals with low occupancies, thus the outer ones).
      I have no idea about the warning "population
 inversion"...
      Since i'm a beginner to this topic, i'd like to ask
 you to summarize to the list.
 Best regards,
 Tamas Karpati
 &&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
 Such warnings can usually be ignored.
 > WARNING:  1 low occupancy (<1.9990e) core orbital  found on  P 3
 >           1 low occupancy (<1.9990e) core orbital  found on  C 5
 In the first case NBO has calculated a couple core orbitals with
 occupancies slightly lower than threshold.  The occupancies reflect
 delocalizing interactions of the two core orbitals with nearby acceptor
 functions.  But the strengths of these interactions are likely so weak
 that they can be neglected.
 Examine the perturbative analysis for interactions involving the two
 orbitals in question.  You'll probably find that the interactions are
 weaker than a kcal/mol or two and can thus be ignored when developing a
 qualitative description of the calculated wavefunction.
 > WARNING:  Population inversion found on atom  P 3
 >           Population inversion found on atom  O 4
 >           Population inversion found on atom  C 5
 >           Population inversion found on atom  C 6
 >           Population inversion found on atom  C 7
 >           Population inversion found on atom  C 8
 >           Population inversion found on atom  C 9
 >           Population inversion found on atom  C10
 >           Population inversion found on atom  O11
 >           Population inversion found on atom  S13
 In the second case NBO has encountered a high energy NAO that has an
 occupancy somewhat greater than a lower energy NAO.  This is a fairly
 common occurence for  extended basis sets for which there are a
 considerable number of Rydberg orbitals identified by NAO.  Examine the
 output for P3, for example.  You might find a 6s NAO with slightly higher
 occupancy than the 5s, or something similar. However, the occupancies of
 these orbitals are probably small enough that they can neglected anyway.
 Eric Glendening
 &&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
 the warning about 'Low Occupancy Core Orbitals' is nothing more than a
 warning with cautious thresholds: just have a look at the occupancies
 to confirm that they are close to 2 electrons.
 I also found the 'Population Inversion' warning in some of my outputs
 and just ignore it if the inversion is between the occupancies of
 lowly occupied orbitals (that is lower than maybe .02 electrons).
 The warning is more common with extended basis sets because then there
 are more unoccupied basis functions and consequently more NAOs with
 near zero occupancies.
 The warning about 'Population Inversion' needs a little background:
 The Natural Population Analysis (NPA) is conducted in the basis of
 Natural Atomic Orbitals (NAOs). They are created by orthogonalizing
 the (atom centered) basis set used in GAUSSIAN under the constraint of
 changing highly populated basis functions as little as possible. On the
 other hand, lowly occupied basis functions may change much more. During
 this process the orbitals of the 'Natural Minimal Basis' (consisting of
 the orbital types that are occupied in the ground state of the atoms,
 e.g. 1s,2s and _all_ 2p orbitals in the case of boron) and the orbitals
 of the 'Natural Rhydberg Basis' (the rest) are treated differently.
 (This has led to some discussion on NPA and transition metals, because
 their np orbitals are treated as Rhydberg Orbitals by the NPA although
 they are sometimes populated, e.g. in complexes of late transition
 metals.) I think that the test for 'Population Inversion' checks if
 the NAO occupancies are decreasing with increasing NAO energy.
 Stefan
 &&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
 Yours,
 Maija Lahtela-Kakkkonen
 ***************************************
 Maija Lahtela-Kakkonen
 Researcher
 CSC-Center for Scientific Computing
 P.O.Box 405
 FIN-02101 ESPOO FINLAND
 TEL 358-9-4572079
 FAX 358-9-4572302
 E-MAIL mlahtela ( ( at ) ) csc.fi
 ***************************************